PROJECT ABSTRACT Sepsis is a leading cause of death, disability, and cost to the healthcare system. Increasing recognition of the devastating burden of sepsis has prompted national performance improvement initiatives and efforts to benchmark the quality of sepsis care. However, efforts to track sepsis incidence, outcomes, and quality have been hindered by the lack of reliable surveillance tools. The candidate, Dr. Chanu Rhee, has thus led recent efforts to develop a novel method for identifying sepsis through electronic health records (EHRs) using objective clinical markers of infection and concurrent organ dysfunction. He has demonstrated that this method is less susceptible to changing diagnosis and coding practices over time than administrative claims data, the traditional method of surveillance, and can be applied to diverse EHR systems in different healthcare systems. The overall goal of this proposal is to build on this work to improve the ability of researchers, clinical organizations, and policy makers to evaluate variation in sepsis incidence, outcomes, and care and thus identify targets for improvement and inform better prevention and treatment strategies. These themes fit with AHRQ's mission to improve health care quality and patient safety, and are reflected in the Specific Aims: 1) Develop and validate a disease severity prediction model for risk-adjustment of sepsis outcomes based on data routinely captured in electronic health records, 2) Evaluate interhospital and intrahospital unit-level variation in sepsis incidence and outcomes using objective EHR-data versus claims codes and identify factors associated with worse outcomes, and 3) Evaluate the impact of lactate testing on clinical outcomes in patients with suspected sepsis. This project will leverage ?big data? with granular clinical information from more than 200 academic and community hospitals and 4 million inpatient encounters, providing the capacity to adjust for multiple confounders, analyze important patient subgroups, and make robust conclusions that are widely generalizable to the nation. The candidate, Dr. Chanu Rhee, MD, MPH, is a junior faculty member in the Department of Population Medicine at Harvard Medical School and a dual-certified critical care and infectious disease physician at Brigham and Women's Hospital. This K08 proposal will capitalize on Dr. Rhee's rich clinical background, foundational training in epidemiology, biostatistics, and public health, and previous research experience with sepsis epidemiology and surveillance using EHR clinical data. Dr. Rhee's immediate goals during this award period are to 1) acquire skills in using ?big data? in healthcare and handling missing data, 2) learn advanced methods in prediction modeling and mixed-effects logistic regression, 3) learn propensity score analysis and comparative effectiveness research skills, 4) improve communication, research leadership, and grant writing skills, and 5) participate in ongoing training in ethical research conduct. These goals will be accomplished through a combination of active mentoring; formal didactic coursework; immersive research training; practical experience; and collaborations with internationally recognized sepsis and infectious disease experts, public health officials, and the Hospital Corporation of America, the largest inpatient private healthcare system in the world. Throughout the K08 award period, he will continue to establish himself as a national thought leader on sepsis epidemiology, surveillance, and quality while preparing for his transition to independence. Dr. Rhee's multidisciplinary and experienced mentorship team, combined with the superb research environment of the Harvard Medical School Department of Population Medicine, provides the support and training experience necessary to achieve his long-term goal of becoming an independently funded health services researcher focusing on using big data to improve disease surveillance, prevention, and quality of care for sepsis and infections in critically ill patients.